Broader 2025–2026 security incidents: zero-days, KEV additions, data breaches, malware campaigns, and supply chain/auto-update compromise outside of explicit AI-agent tradecraft

General Zero-Days, Breaches & Supply Chain

The cybersecurity landscape entering 2027 remains dire and increasingly complex, marked by a relentless escalation in zero-day exploits, expanding Known Exploited Vulnerabilities (KEV) catalogs, pervasive supply chain compromises, and sprawling data breaches. Recent developments throughout late 2026 and early 2027 not only reinforce earlier trends but also introduce new dimensions of threat activity, particularly involving sophisticated ransomware extortion campaigns, IoT and firmware persistence mechanisms, and the weaponization of trusted tooling and CI/CD pipelines. Collectively, these incidents underscore an urgent imperative for adaptive, cryptographically assured, and AI-native defenses to safeguard critical infrastructure, enterprise ecosystems, and sensitive data.

Continued Surge in Zero-Day Exploitation and KEV Expansion

The frequency and sophistication of zero-day attacks have accelerated into 2027, confirming that adversaries are increasingly targeting orchestration platforms, CI/CD toolchains, backup systems, and firmware layers as high-value attack surfaces:

Google’s Threat Analysis Group (GTIG) reports stealthy zero-day exploits remain a staple for commercial spyware vendors and advanced persistent threat (APT) groups, particularly against enterprise and critical infrastructure targets. Notable among these are vulnerabilities in Cisco SD-WAN (CVE-2026-20127) and TeamCity (CVE-2026-28194), both actively exploited to compromise network and build automation environments.
The CISA KEV catalog continues to swell, with additions focused on orchestration tools such as SolarWinds, Ivanti Endpoint Manager, Dell RecoverPoint, and n8n automation platforms. These flaws facilitate lateral movement and supply chain contamination, highlighting the criticality of patch prioritization.
Patch cycles remain intense. Microsoft’s March Patch Tuesday cycles in 2026 and 2027 addressed upwards of 80 vulnerabilities at a time, including zero-days in Windows, Office, SQL Server, and embedded AI runtimes. Early 2028 saw urgent patches from Veeam for seven critical remote code execution vulnerabilities in backup and replication infrastructure, underscoring the growing acknowledgment of backup systems as prime attack vectors.
Firmware-level vulnerabilities patched in Qualcomm and MediaTek chipsets via Google’s March 2026 Android update have been linked to persistent malware capable of bypassing hardware security modules. These sophisticated implants raise alarms about long-term persistence threats to mobile devices and IoT ecosystems.

Escalating Supply Chain Intrusions and CI/CD Pipeline Poisoning

Supply chain attacks have deepened in complexity and scope, exploiting developer trust and automation workflows:

The ForceMemo campaign, uncovered in 2028 by StepSecurity, compromised hundreds of Python repositories on GitHub through malicious transitive dependency injections. This tactic mirrors and expands upon earlier campaigns like GlassWorm, which weaponized Open VSX extensions, demonstrating attackers’ growing proficiency in dependency poisoning to infiltrate software supply chains silently.
Exploitation of the TeamCity open redirect vulnerability enabled attackers to inject malicious artifacts into CI/CD build processes, illustrating how automation pipelines—once deemed secure internal assets—are now leveraged as stealthy footholds for lateral movement and software supply chain compromise.
Vulnerabilities such as Chartbrew’s CVE-2026-25887 in orchestration dashboards grant persistent privileged access within development environments, amplifying supply chain risks.
FBI operations have highlighted the operational impact of supply chain compromises, including takedowns related to the LeakBase data trading platform and breaches affecting suppliers within FBI wiretap networks, emphasizing the critical need to secure upstream dependencies and vendor ecosystems.

Signed Malware and Trojanized Tooling: Abusing Trust at Scale

Adversaries are refining stealth techniques by leveraging legitimate trust mechanisms and software authenticity:

The TrustConnect threat group has been observed distributing cryptographically signed malware binaries masquerading as legitimate workplace applications such as Microsoft Teams and Zoom. These signed binaries deploy weaponized Remote Monitoring and Management (RMM) tools, granting persistent, covert access that bypasses traditional detection mechanisms by exploiting trust in digital signatures.
The ongoing risks of trojanized AI coding assistants were exemplified by the McKinsey Lilli/CodeWall breach, where vendor trust was exploited without robust cryptographic attestation. Fake installer guides for AI tools like Claude Code have been weaponized to spread credential stealers and backdoors, threatening the integrity of AI development workflows.
The market for Windows zero-day remote code execution exploits remains inflated, with prices recently surpassing $220,000 on dark web forums. This pricing pressure compresses defensive patch windows and empowers attackers with advanced capabilities.

Large-Scale Data Breaches and Ransomware Extortion Campaigns

Data breaches continue to proliferate with severe consequences for privacy and enterprise trust, while extortion campaigns grow more aggressive and targeted:

The Conduent ransomware attack revealed in late 2026 impacted tens of millions across multiple U.S. states, exposing 25 million medical and Social Security records. This incident highlights the vulnerability of government contractors handling sensitive citizen data.
The CarGurus breach, linked to the ransomware group ShinyHunters, exposed 12.4 million records, including personal information from users who searched for vehicles on the platform. ShinyHunters has leveraged these breaches to escalate extortion efforts against cloud CRM platforms such as Salesforce Aura by exploiting configuration flaws and leaked data.
The Panera Bread data leak exposed customer information including names, emails, physical addresses, birthdays, and partial credit card numbers. Notably, Panera allegedly ignored warnings about the data exposure, underscoring ongoing failures in breach response protocols.
Other notable breaches include the LexisNexis incident, which compromised 3.9 million records due to an unpatched React vulnerability, and an AWS breach exposing 400,000 cloud user profiles, both highlighting persistent security gaps in web frameworks and cloud configurations.
The Divine Skins breach (March 2026) and the leak of 60,000 “Heart DATA” records from Indonesia’s Buol Regency Government website (buolkab.go.id) demonstrate that government and niche sectors remain highly vulnerable to data exposure.
Leaked data from the U.S. Department of Homeland Security’s technology incubator has exposed sensitive AI surveillance research funding and initiatives, raising concerns about operational security and geopolitical ramifications if such data is weaponized.

IoT and Firmware Persistence: Long-Term Compromise at Scale

Firmware-level compromises and IoT botnets are increasingly leveraged for persistence and AI-driven attacks:

The Kadnap router botnet, infecting over 14,000 devices worldwide, exploits hardware flaws in MediaTek and Qualcomm chipsets to implant persistent firmware malware. This malware evades conventional endpoint detection and serves as a platform for AI-driven lateral movement and malware hosting.
In 2027, medical device manufacturer Stryker fell victim to a cyberattack attributed to Iranian threat actors, spotlighting the grave risks firmware and supply chain attacks pose to life-critical healthcare infrastructure.
New attack vectors involving zero-click cross-site scripting (XSS) against AI chatbots and browser extensions have emerged, enabling session hijacking and data exfiltration without user interaction. These exploits combine postMessage misconfigurations with prompt injection and sandbox escape techniques.
The FalkeAI Browser Agent demonstrated advanced prompt injection attacks that implant covert commands inside AI sandboxes, circumventing traditional detection and raising the bar for AI sandbox security.

Defensive Imperatives: Toward Adaptive, Multi-Layered Security

In response to this evolving and multifaceted threat environment, organizations must urgently adopt integrated defense strategies:

Proof-Based Vendor Assurance: Enforcing cryptographic attestation, strong installer signing, and continuous runtime integrity monitoring is essential to counter trojanized tooling and transitive dependency poisoning campaigns like ForceMemo and GlassWorm.
CI/CD Pipeline and Workflow Hardening: Implementing rigorous controls over build workflows, dependency vetting, and artifact provenance helps prevent injection attacks and open redirect vulnerabilities that compromise software supply chains.
Zero-Trust Micro-Segmentation: Segmenting networks around orchestration consoles (e.g., n8n, Chartbrew), backup infrastructure (Veeam), and IoT/OT devices limits lateral movement and reduces breach blast radius.
AI-Native Detection and Response: Deploying AI-enhanced behavioral anomaly detection and context-aware vulnerability scanners (e.g., OpenAI Codex Security) is critical for identifying stealthy autonomous malware campaigns, prompt injection, and sandbox escape exploits.
Firmware and IoT Security: Fortifying hardware security modules and enforcing supply chain integrity validation are crucial to prevent persistent firmware compromises and botnet formation.
Credential Governance and MFA: Adopting hardware-backed, phishing-resistant multi-factor authentication alongside frequent credential rotation and vigilant monitoring remains a cornerstone defense against credential theft propagated via trojanized installers and signed malware.

Outlook and Implications

The cybersecurity epoch spanning 2025 through early 2027 reveals an adversarial ecosystem that is not only expanding in scale but also innovating in stealth, persistence, and impact. The convergence of zero-day exploitation, KEV catalog expansion, supply chain poisoning, signed malware distribution, and large-scale data breaches—including high-profile incidents affecting government contractors and consumer platforms—signals that no sector is immune.

The amplification of IoT/firmware persistence threats and AI-targeted sandbox escapes further complicates defense efforts, requiring organizations to adopt holistic strategies that integrate cryptographic proof, rigorous CI/CD hardening, zero-trust micro-segmentation, AI-native detection, and robust credential governance.

Without accelerated adoption of these layered defenses and a heightened sense of patch urgency, the velocity and complexity of adversary tradecraft risk overwhelming existing security postures. This could result in not only catastrophic breaches but also a profound erosion of trust in digital ecosystems worldwide.

Selected References

Google GTIG: 90 zero-day flaws exploited in 2025 as enterprise targets grow
Microsoft March Patch Tuesday: 83 vulnerabilities fixed including zero-days
CISA Adds SolarWinds, Ivanti, Dell RecoverPoint, and n8n Flaws to KEV Catalog
ForceMemo: Hundreds of GitHub Python Repos Compromised via Transitive Dependency Injection
Signed Malware Impersonating Workplace Apps Used To Deploy RMM Backdoors
TeamCity Open Redirect Vulnerability (CVE-2026-28194) Threatens CI/CD Pipelines
Veeam Patches 7 Critical Backup & Replication Flaws Allowing Remote Code Execution
Kadnap Malware Infects 14,000 Routers Worldwide
LexisNexis Hack Exposes 3.9M Records Through Unpatched React Vulnerability
Salesforce Sounds Alarm Over Fresh Data Extortion Campaign
FalkeAI Browser Agent Demonstrates Advanced Prompt Injection
Windows Zero-Day Exploit Market Surges to $220,000
Stryker Cyberattack Linked to Iranian Hacker Group
Critical Alert: Recent Divine Skins Data Breach (2026)
60,000 “Heart DATA” Records of Buolkab.go.id Allegedly Leaked
Hacked Data Shines Light on Homeland Security’s AI Surveillance Ambitions
Conduent Data Breach Hits Tens of Millions Across U.S. States
CarGurus Breach Linked to ShinyHunters Exposes 12.4M Records
Panera Bread Ignored Customer Data Leak Warnings
Cyber Security Tribe Releases 2026 Annual State of the Industry Report

The evolving adversary landscape demands relentless innovation and operational rigor. Organizations must prioritize securing software supply chains, enforcing patch discipline, fortifying firmware integrity, and integrating AI-native detection to stay ahead of increasingly autonomous and sophisticated threat actors.

Sources (153)